- From: Ian Hickson <ian@hixie.ch>
- Date: Mon, 23 Feb 2009 08:05:06 +0000 (UTC)
- To: David Woolley <forums@david-woolley.me.uk>
- Cc: Philip Taylor <pjt47@cam.ac.uk>, www-html@w3.org, Giovanni Campagna <scampa.giovanni@gmail.com>
On Mon, 23 Feb 2009, David Woolley wrote: > > I thought it was one of the key design principles of HTML5 that all > parsers treat parsing "errors" alike. That means that, if a DOM based > parser can recover from an error, all conforming parsers must do so. No, the key design principle is that if UAs want to recover from errors, they be required to do so in a well-defined way. However, nothing stops a tool that parses HTML5 from refusing to handle invalid content, and the spec explicitly allows this: http://www.whatwg.org/specs/web-apps/current-work/multipage/parsing.html#parse-error HTH, -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 23 February 2009 08:08:08 UTC